Beyond severity: Mapping cognitive heterogeneity in schizophrenia at the structural level

Abstract

Cognitive heterogeneity is a core feature of schizophrenia (SCZ), but it remains unclear whether this heterogeneity reflects graded deviation within a shared cognitive structure or alterations in that structure. In this study, latent cognitive structure was delineated in healthy controls (HC) and evaluated in SCZ using a norm-anchored cognitive structural deviation (NCSD) framework. A stable single-factor structure was identified in HC, indicating that diverse cognitive measures were captured within a common latent structure. Compared with HC, patients with SCZ showed greater overall deviation from the normative structure together with altered loading patterns in WCST-derived executive indicators, suggesting both magnitude differences and structural alteration. The global deviation index demonstrated stable performance in differentiating SCZ from HC under cross-validation. Structural deviation indices were associated with anxiety ratings and showed trend-level associations with positive symptoms. Exploratory clustering further suggested that cognitive heterogeneity in SCZ may reflect distinct profiles characterized by different combinations of deviation magnitude and loading alteration. NCSD provides a refined framework for characterizing cognitive heterogeneity beyond approaches that consider deviation severity or structural variation separately. Further studies will be needed to establish the robustness and broader applicability of this framework.